Back

Literature-based predictions of treatments for genetic disease pathology

Deisseroth, C. A.; Lee, W.-S.; Kim, J.-Y.; Jeong, H.-H.; Wang, J.; Zoghbi, H. Y.; Liu, Z.

2022-09-12 bioinformatics
10.1101/2022.09.08.506253 bioRxiv
Show abstract

Identifying genetic modifiers of disease-causing genes can guide drug discovery for treatment of genetic disorders, but selecting promising drugs to test requires extensive and up-to-date knowledge of drug-gene and gene-gene relationships. To address this challenge, we present PARsing ModifiErS via Abstract aNnotations (PARMESAN), a computational tool that searches PubMed for information on these relationships, and assembles them into one knowledgebase. PARMESAN then hypothesizes on undiscovered drug-gene relationships, assigning an evidence-based score to each hypothesis. We compare PARMESANs drug-gene hypotheses to all of the drug-gene relationships displayed by DrugBank, and see a strong correlation between the prediction score and the predictive accuracy--such that predictions scoring above 10 are 11 times more likely to be correct than incorrect. This publicly available tool provides an automated way to prioritize drug screens to target the most-promising drugs to test, thereby saving time and resources in the development of therapeutics for genetic disorders.

Matching journals

The top 12 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.